Evaluations
List benchmark catalog
GET
List benchmark catalog
Authorizations
API key with the vk_ prefix. Create one from the dashboard.
Response
200 - application/json
Available benchmark tasks
Catalog id used as task in a benchmark scorer (e.g. "gsm8k").
"inference_only" or "code_exec" (code_exec runs each sample in a sandbox).
Total rows in the source split; cap with sample_limit.
List benchmark catalog

